AEWDAT - African Easterly Waves: Identified and Tracked Wave Troughs

This data set contains identified wave troughs within African Easterly Waves, utilizing 700 hPa curvature vorticity data from the ERA5 reanalysis spanning from June to October over the period 1940-2022. It encompasses all identified wave troughs and also includes extracted tracks derived from these identified features.

The identification methodology closely follows the approach outlined by Belanger et al. (2016). Troughs and ridges are determined as zeros in Curvature Vorticity Advection fields. Threshold criteria, based on a climatology and the second derivative, are applied to extract the troughs. Tracking is accomplished through a heuristic that connects wave troughs meeting specified criteria for speed and average speed over a 36-hour period.

The JSON files within this dataset provide a description of the wave troughs, structured as follows:

1. Each file contains one year of data, with one set each.
2. Each set includes:
   • A list of timestamps, where each timestamp contains the identified troughs at a specific time. Each time entry contains a list of objects (wave troughs), with each containing a set-unique identifier and properties describing the wave trough, such as its bounding box and line string representation.
   • The data graph, consisting of a list of edges, wherein each edge connects one parent object to multiple child objects, forming the tracking graph. A connection between a parent and child indicates that they represent the same wave trough entity, with the child being the next manifestation of the parent entity along the time line.
   • A list of tracks, with each track represented as a data graph as well.

Furthermore, a zip-file (wtsnc.zip) contains netCDF representations of the wavetroughs as labeled data field, where the labels correspond to the identifier in the JSON file. The file contains fields for all wavetroughs, only wavetroughs making up the tracks, and the tracks labeled by track ID. Be careful when extracting, the data is heavily compressed.

Further information regarding this strategy can be found in an upcoming publication, soon to be submitted for review.